What the Are You You? Game Shows About Face Recognition
Updated
Are You You? (areyouyou.eu) is a browser-based game that shows how a face recognition algorithm turns your face into a match score — and how hard it is to push that score down. You win by making faces until the match meter drops below 50%. The takeaway: disguising your face can lower an algorithm's confidence, but the game is built to demonstrate how difficult it is to fully escape face recognition, not to prove you can.
What the game actually measures
The game runs a face recognition model locally in your browser. It compares your live camera image against a reference photo and outputs a percentage match — how likely the algorithm thinks the two images show the same person.
That percentage is the whole point. Face recognition isn't a yes/no verdict; it's a confidence score. The game makes that visible:
- A high match percentage means the algorithm is confident it's you.
- Below 50% counts as fooling it — you win the round.
- Your results screen shows your score, your average match percentage, and the photos where you looked least like yourself to the algorithm.
Why making faces can beat the algorithm
The game's own tips list what tends to work: move your head side to side and up and down, stick out your tongue, laugh, remove glasses, and use your hands to deform your face. These all change the geometry and features the model relies on — head pose, mouth shape, eye region, face outline.
The flip side is the lesson the project is making: you can lower confidence with deliberate distortion, but you can't reliably disappear. A determined system with multiple frames, better angles, or other signals would still have material to work with. The game is a light-hearted demonstration of that gap, not a privacy tool.
What happens to your camera data
According to the site, camera access is required to play, but no personal data is collected. The face recognition algorithm runs entirely in your browser, no photos are uploaded anywhere, and there are no cookies or third-party tracking. The only thing stored is your final score, which the site says acts as a viewer count.
If you're evaluating the game for a classroom or workshop, that local-only design is the relevant detail: the privacy claim is specific and checkable in how the page behaves, and the project is transparent about its purpose.
Who made it and why
Are You You? is an educational project by artist and privacy designer Tijmen Schep, made as part of the Sherpa project — an EU-funded research consortium exploring how AI systems could better protect human rights. The stated goal is to show how difficult it can be to escape face recognition technology, in a light-hearted way. The game carries the #ReclaimYourFace tag and points visitors toward related privacy advocacy.
How to play, step by step
- Open areyouyou.eu and allow camera access when prompted — the game states this is necessary and that nothing is uploaded.
- Wait for the face recognition model to load, then look at the camera; your photo is taken after a 5-second countdown.
- Make faces using the suggested moves until the match meter falls below 50%.
- Review your results: total score across levels, your average match percentage, and the photos that looked least like you.
- Optionally download your result picture and share with the privacy-friendly buttons (the site says these contain no third-party tracking).
Common snags: if the camera can't be accessed, the game shows an "Unable to access camera" message — check browser camera permissions. Lighting and distance from the camera affect how consistently the model reads your face, so results vary between attempts.
What to take away
The game is best read as a demonstration of two things at once: face recognition reduces a person to a similarity score, and that score can be manipulated at the edges but not switched off by mugging for the camera. For anyone teaching or learning about facial recognition, it's a concrete, hands-on way to see the confidence number move — and to understand why "just make a funny face" isn't a real defense.